Re: [Axiom-developer] Sockets, threads and GUI<->Kernel communication

[Gabriel Dos Reis]

C Y <address@hidden> writes:

| --- Gabriel Dos Reis <address@hidden> wrote:
| 
| > | I do agree that threads are a more attractive way to implement the
| > | client-server setup when dealing with a strictly local copy of
| > | Axiom,
| > 
| > If you agree with that, then you must agree with the over-engineering
| > characterization.  We have more uses of Axiom as local copy than we
| > have as a truly distributed system over network.
| 
| At the moment yes, but the future is unknown.

If you can't handle the known present, how do you plan to handle the
unknown future?

|  It is far easier to
| design the system from the beginning to be flexible (client-server) and
| then not use that ability.

Please, note that "client-server" does not necessarily imply sockets,
and using threads does not necessarily exclude "client-server".

However, my fundamental point remains: common things should be simple.

I think we should be careful about the arguments we make.

[...]

| > A 101 principle of engineering is to keep common thing simple.
| 
| Sockets seem to be a minor detail to me (not technically, perhaps) -

Yet, they are they roots of many unpleasant aspects when working on
the internal details and improvements of Axiom.

[...]

| broadly developed and robust.  That may have changed since, but I would
| want more information than I have now before I would be willing to
| regard thread support in Lisp as something to depend on.

Well, my initial comment is more a statement about the current state
as opposed to an active suggestion to redesign Axiom.  Certainly, we
have around Lisp systems mapping to threads and using them.  They
don't look more complicated than the current mess we have.  Now, you
may have a point about standard thread interface, but that is more of
a general problem with Lisp environments once you start doing anything
useful -- you have to catter for all kinds of variabilities.

| > | Allow Axiom to have both local and remote interface clients. 
| > 
| > Well obviously, no improvement will remove the socket support.
| > The point wasn't that socket are useless -- quite the contrary.  But,
| > the engineering has forgotten the basics.
| 
| Thread support in lisp implementations is not usually described as
| basic, in my experience... maybe to use but certainly not to implement.
| 
| Do you mean the CURRENT implementation of things in Axiom is not well
| engineered, irrespective of the use of sockets?

Yes.

| > I'm not a windows developer, so patches are welcome.
| 
| I'll take a crack at NSIS once I have a Windows box available.
| 
| Cheers,
| CY
| 
| P.S.  I don't mean to be an irritation, Gaby - it's possible I'm just
| lacking enough basic understanding to follow you correctly.  Is there
| some introductory material on thread vs. socket issues that I could read?

They are not mutually exclusive, so I don't understand your point.

-- Gaby